3Com Megahertz 3CXM756 PCMCIA modem on 5.3

[Astrodog]

Sorry. Ended up sending it only to Duane. Oops!


---------- Forwarded message ----------
From: Astrodog <astrodog at gmail.com>
Date: Sun, 30 Jan 2005 15:55:31 -0800
Subject: Re: 3Com Megahertz 3CXM756 PCMCIA modem on 5.3
To: Duane Winner <dwinner-lists at att.net>


On Sun, 30 Jan 2005 17:44:48 -0500, Duane Winner <dwinner-lists at att.net> wrote:
> Astrodog wrote:
>
> >On Sun, 30 Jan 2005 12:26:50 -0500, Duane Winner <dwinner-lists at att.net> wrote:
> >
> >
> >>M. Warner Losh wrote:
> >>
> >>
> >>
> >>>In message: <41F6C64B.3040700 at att.net>
> >>>           Duane Winner <duanewinner at att.net> writes:
> >>>: Warner Losh wrote:
> >>>:
> >>>: >>Thanks! That would be great. At least I know it's not just me.
> >>>: >>
> >>>: >>
> >>>: >
> >>>: >I've updated what I think is the fix to RELENG_5, so please test and
> >>>: >let me know.
> >>>: >
> >>>: >
> >>>: >
> >>>: That did the trick. I took one of my other T30's and changed my cvs tag
> >>>: from RELENG_5_3 to RELENG_5, cvsup'd, buildword and buildkernel, and I
> >>>: was able  to dial out.
> >>>:
> >>>: I get this now when inserting the card:
> >>>:
> >>>: pccard1: Allocation failed for cfe 32
> >>>: sio4: <3Com Megahertz 3CXM756/3CCM756> at port 0x2f8-0x2ff irq 5
> >>>: function 0 config 33 on pccard1
> >>>: sio4: type 16550A
> >>>: sio4: unable to activate interrupt in fast mode - using normal mode
> >>>:
> >>>: And the obligatory 'detach' when I remove it.
> >>>
> >>>Cool.  Sounds about right.  Not sure you should be getting the
> >>>allocation failed bit, but that's harmless, as is the fast interrupt
> >>>message.
> >>>
> >>>: Any chance of this getting into RELENG_5_3? I'm not sure how that works
> >>>: (since 5.3 gets "security and critical fixes" only), but I thought I
> >>>: would ask. Does this classify as a critical fix?
> >>>
> >>>Hmmm.  Lemme ask.
> >>>
> >>>Warner
> >>>
> >>>
> >>>
> >>>
> >>Hi Warner,
> >>
> >>I was just wondering what the status of this is, if any. I'm about
> >>commit our in-house installation procedures for 5.3, and others on my
> >>team are going to want to migrate soon, but lack of PPP support on our
> >>laptops isn't going to fly over too well, especially now that there is
> >>some travelling going on. We want to stick to -release, as opposed to
> >>-stable.
> >>
> >>I was thinking, I figure we have three possible options:
> >>
> >>1. If your patch is going to get merged into 5.3-release, that would be
> >>the best route; but if that isn't going to happen, or not soon:
> >>
> >>2. Could we manually apply your patch to /usr/src after cvsup'ing before
> >>buildword/buildkernel?
> >>
> >>3. Find a different PCMCIA modem that is supported by 5.3-release. Most
> >>seem to be dirt-cheap these days, so if there is another one that works,
> >>we wouldn't be opposed to picking these up.
> >>
> >>Your thoughts/advice?
> >>
> >>Cheers,
> >>Duane
> >>
> >>
> >>
> >>
> >>>
> >>>
> >>_______________________________________________
> >>freebsd-mobile at freebsd.org mailing list
> >>http://lists.freebsd.org/mailman/listinfo/freebsd-mobile
> >>To unsubscribe, send any mail to "freebsd-mobile-unsubscribe at freebsd.org"
> >>
> >>
> >>
> >
> >If its in RELENG_5 (-STABLE) you oughta be able to cvsup to that. Imo,
> >with mobiles, for the time being they should be running -STABLE if
> >they have support problems anyway.
> >
> >
> Actually, I have considered that as a possibility.
>
> What are the most common negatives of running stable instead of release?
> I have no experience tracking -STABLE, as we have always tracked
> -RELEASE only.
>
> I think most of our obsession with running -RELEASE is more historical
> than anything else, and the impression we've had since we started using
> FreeBSD is that -RELEASE is more 'stable' then -STABLE.
>
> Also, I'm fairly certain that when we start upgrading servers, we
> definately want to to go -RELEASE. The software developers on my team
> generally try to maintain their laptops (they telecommute from far away,
> so their laptops are more or less their 'lab' environments) as close to
> as possible to my server configurations, so that there are no surprises
> and we can generally apply the same knowledge between the two.
>
> Maybe I'm just being more anal-retentive than I need to be. :)
>
> Cheers,
> Duane
>
> >
> >
>

-STABLE is generally less stable than -RELEASE, but you won't see
fixes, like these backported to -RELEASE, its for security stuff only
from my understanding. -STABLE is slightly more administrative
overhead, since you need to watch the lists, but personally, I've had
great luck with it. On a few occassions, we've actually had to deploy
-CURRENT in production, which worked out alright... but is a nightmare
from an administrative perspective. Until 5.4 comes out, I suspect
I'll be tracking -STABLE on all of my boxes, while they work out the
last few things.

--- Harrison Grundy